diff --git a/pipeline_cogvideox.py b/pipeline_cogvideox.py
index fefd0bc..ae1aa6d 100644
--- a/pipeline_cogvideox.py
+++ b/pipeline_cogvideox.py
@@ -536,11 +536,10 @@ class CogVideoXPipeline(VideoSysPipeline, CogVideoXLoraLoaderMixin):
             logger.info(f"latents: {latents.shape}")
             logger.info(f"mask: {mask.shape}")
 
-        # 7. Denoising loop
+      
         num_warmup_steps = max(len(timesteps) - num_inference_steps * self.scheduler.order, 0)
-        comfy_pbar = ProgressBar(num_inference_steps)
 
-        # 8. context schedule and temporal tiling
+        # 7. context schedule and temporal tiling
         if context_schedule is not None and context_schedule == "temporal_tiling":
             t_tile_length = context_frames
             t_tile_overlap = context_overlap
@@ -560,7 +559,7 @@ class CogVideoXPipeline(VideoSysPipeline, CogVideoXLoraLoaderMixin):
             use_temporal_tiling = False
             use_context_schedule = False
             logger.info("Temporal tiling and context schedule disabled")
-            # 8.5. Create rotary embeds if required
+            # 7.5. Create rotary embeds if required
             image_rotary_emb = (
                 self._prepare_rotary_positional_embeddings(height, width, latents.size(1), device)
                 if self.transformer.config.use_rotary_positional_embeddings
@@ -569,7 +568,7 @@ class CogVideoXPipeline(VideoSysPipeline, CogVideoXLoraLoaderMixin):
             if tora is not None and do_classifier_free_guidance:
                 video_flow_features = tora["video_flow_features"].repeat(1, 2, 1, 1, 1).contiguous()
 
-        # 9. Controlnet
+        #8. Controlnet
         if controlnet is not None:
             self.controlnet = controlnet["control_model"].to(device)
             if self.transformer.dtype == torch.float8_e4m3fn:
@@ -604,8 +603,9 @@ class CogVideoXPipeline(VideoSysPipeline, CogVideoXLoraLoaderMixin):
                 for param in module.parameters():
                     param.data = param.data.to(device)
 
-        # 10. Denoising loop
-        with self.progress_bar(total=num_inference_steps) as progress_bar:    
+        # 9. Denoising loop
+        comfy_pbar = ProgressBar(len(timesteps))
+        with self.progress_bar(total=len(timesteps)) as progress_bar:    
             old_pred_original_sample = None # for DPM-solver++
             for i, t in enumerate(timesteps):
                 if self.interrupt: